Relevant news re malfunctioning AF under artificial light

Tommot1965 · 02-10-2011, 03:23 AM

Originally posted by simico

Pffff....

LOL

jolepp · 02-10-2011, 03:48 AM

It just occurred to me that EV alone might not be a sufficient measure of problematic conditions, in addition the max. aperture of the lens used is needed as that is used for AF: the amount of light (intensity) on the AF sensor depends on these two. We might want to use an equivalent value for some arbitrary f-stop, e.g. Christine's figures of EV 2.8 and EV 2.2 with f=1.9 are equivalent to EV 3.9 and EV 3.3 for f=2.8. (The adjustment term comes from the ratio of aperture areas to which light intensity is directly proportional: 1/1.9^2 / 1/2.8^2 = (2.8/1.9)^2 = 2.1717... log2 of this gives the difference in EV units as 1.1.) The suggestion of "f=2.8 normalized" figures is implied.

falconeye · 02-10-2011, 05:48 AM

Originally posted by Christine Tham

Forgot to mention, I had a hard time repeating the test, because the AF assist light keep coming on, even at EV3-EV4. Of course, once the AF assist light comes on, the camera is able to focus with no FF.

Christine, you probably missed my earlier question:

Did the AF assist light (in your camera) assist the entire focus process or just the beginning before the lens is shifted (it makes a difference if the lens wasn't already in focus).

Mine uses AF assist only for the first measurement (if at all), not thru hunting. Meaning it would still miss the focus if it wasn't in focus initially. I.e., it does fine focus corrections in the dark even if starting with AF assist light on.

areidjr · 02-10-2011, 06:12 AM

Originally posted by falconeye

Did the AF assist light (in your camera) assist the entire focus process or just the beginning before the lens is shifted (it makes a difference if the lens wasn't already in focus).

Mine uses AF assist only for the first measurement (if at all), not thru hunting. Meaning it would still miss the focus if it wasn't in focus initially. I.e., it does fine focus corrections in the dark even if starting with AF assist light on.

FWIW and if you missed my earlier reply, mine behaves exactly as your. Although, on rare occasions it seem to work as expected.

falconeye · 02-10-2011, 09:34 AM

Originally posted by areidjr

FWIW and if you missed my earlier reply, mine behaves exactly as your. Although, on rare occasions it seem to work as expected.

Thanks.
But Christine claims to nail focus if AF assist comes on. So, I'd like to see her response as well

**GregK8** · 02-10-2011, 10:38 AM

If that all it is, why not allow the Fx button to be programmed to turn on the AF assist light?

Christine Tham · 02-10-2011, 11:03 AM

Originally posted by falconeye

Christine, you probably missed my earlier question:

Did the AF assist light (in your camera) assist the entire focus process or just the beginning before the lens is shifted (it makes a difference if the lens wasn't already in focus).

Mine uses AF assist only for the first measurement (if at all), not thru hunting. Meaning it would still miss the focus if it wasn't in focus initially. I.e., it does fine focus corrections in the dark even if starting with AF assist light on.

Hi Falk

Sorry I missed out on your earlier post - I tuned out of the thread and wasn't reading the posts.

In response to your query: On my K-5, *if* the AF assist light comes on, it stays on until the camera locks focus, and the end result is a perfectly focused photo.

That's why I had to ensure the AF assist light was not engaging when doing the test.

PS - love the work you did on sensor blur when SR is on

Christine Tham · 02-10-2011, 11:10 AM

Originally posted by jolepp

It just occurred to me that EV alone might not be a sufficient measure of problematic conditions, in addition the max. aperture of the lens used is needed as that is used for AF: the amount of light (intensity) on the AF sensor depends on these two. We might want to use an equivalent value for some arbitrary f-stop, e.g. Christine's figures of EV 2.8 and EV 2.2 with f=1.9 are equivalent to EV 3.9 and EV 3.3 for f=2.8. (The adjustment term comes from the ratio of aperture areas to which light intensity is directly proportional: 1/1.9^2 / 1/2.8^2 = (2.8/1.9)^2 = 2.1717... log2 of this gives the difference in EV units as 1.1.) The suggestion of "f=2.8 normalized" figures is implied.

Interesting.

However, I can't seem to replicate the FF on the DA 70mm 2.4, even right down to EV 1.

I also can't replicate the FF on a cheap lens - FA 28-105mm (at f3.2).

I haven't tried my FA 24 at f2 but I am not sure the depth of field would be deep enough either.

The error in focusing appear to be less than 5mm.

Gimbal · 02-10-2011, 12:14 PM

At which distance is it best to test for FF? (Mine is so severe that I can aim at anything and it will be fuzzy unless I use live view.) I notice that many tests at near minimum range, is that the best distance to test it? I know that the DOF will be shortest but then the focusing precision will be at it's highest as well.

I find 2-3meters to be really revealing om my camera. First focus using live view, then without moving the camera activate phase focus and watch the focus ring, it shouldn't move.

raphtze · 02-10-2011, 12:36 PM

Originally posted by philbaum

I've seen a coupla identical posts on the Nikon D7000 dslr forum on dpreview, where a camera had backfocus in normal daylight that shifted to FF in tungsten low light. Its hard for me to conceive that a mfr would ship a camera this way, but some of the similarities of this issue between Nikon and Pentax are certainly interesting.

i am sure both companies wanted to get their latest and greatest pro-sumer camera out to maximize profit....but well....as we're now seeing, it probably wasn't the best. some what tangential....i wonder what the QC and testing is like when a sensor is finally developed, i would assume that pentax and nikon would be a lot more careful--well, perhaps that is what it will be like with D7K and K-5 successors. hehe one thing i know is, us forum members would probably be hired on the spot for such testing

-raph

Smeggypants · 02-10-2011, 01:47 PM

Originally posted by GregK8

If that all it is, why not allow the Fx button to be programmed to turn on the AF assist light?

My K20D doesn't need an assist light to focus right down to low light levels.

**jpzk** · 02-10-2011, 01:48 PM

So ... I did some AF tests using the suggested parameters by another member:

ISO 3200
1/15 to 1/25s
F1.8 - - lens is a FA31mm Ltd , also with the DA*16-50 at F2.8
Distance to subject: 18-24 inches
Tungsten lighting, dimmed to achieve the above parameters.

Results: FF with both set ups (FA31 Ltd and DA*16-50).
I even got FF at ISO1250, with the same shutter speeds and apertures.

Whether this is important or not, I had the AF assist light come on a couple of times and that did not make any difference.
I tried to expose the subject so that the assist lamp would not come on, but it did, as I said, a couple of times (only).

While those are not parameters I would likely use in the "real world", I was disappointed to find out that my K5 does show this low light (Tungsten) AF issue.
[I can "live with that" though but am I the only one who will? Just wondering.]

I contacted Pentax Imaging first, and ended up in the USA, which is fine. The fellow was trying to be helpful and mentioned that there have been a few reports of such AF issues.
I asked whether they had heard anything at all regarding a fix of some sort (either firmware or hardware/return the camera ... blah, blah, blah ...) and he said "No". He also said the there has been no report from Pentax Japan, so they would not have a clue on how this is going to be handled.
He suggested to contact Pentax Canada because I bought the K5 there.

I called Pentax Canada: of course, they are as clueless as usual; the fellow at the repair/service department told me that I am the first person reporting such problem! No kidding!!

So I suppose this is a matter of waiting for Pentax Japan for anyhting they would come up with to fix this, and perhaps I am being over confident that they will?

I haven't posted any pics from the tests (yet) ... they are still on the SD card. Eventually, I will post them.

For what it was worth, this was my contribution for the issue at hand.

Cheers!

JP

areidjr · 02-10-2011, 02:06 PM

Originally posted by jpzk

I contacted Pentax Imaging first, and ended up in the USA, which is fine. The fellow was trying to be helpful and mentioned that there have been a few reports of such AF issues.
I asked whether they had heard anything at all regarding a fix of some sort (either firmware or hardware/return the camera ... blah, blah, blah ...) and he said "No". He also said the there has been no report from Pentax Japan, so they would not have a clue on how this is going to be handled.

This is totally consistent with the calls I made to Pentax USA both a little over 2 weeks ago and again yesterday.

falconeye · 02-10-2011, 03:30 PM

Originally posted by Christine Tham

In response to your query: On my K-5, *if* the AF assist light comes on, it stays on until the camera locks focus, and the end result is a perfectly focused photo.

Now, that's interesting. Let's double check our firmware versions first, though.

Unfortunately, there has been a silent firmware upgrade recently (which I did Not install). I'll post the exact release and checksum, and how you can compute your checksum asap. What OS (Mac, PC or Linux) do you use? Maybe, Pentax silently changed the way AF assist light behaves.

Christine Tham · 02-10-2011, 03:41 PM

Kind of off topic - but over the years I used many different cameras, and in my experience most of them struggle to focus in low light.

If you recall, I mentioned that under the conditions in which my K-5 will FF, the Canon EOS-1D Mk III refuses to lock focus.

And if you like watching videos, here's an example of front focusing under low light from my Sony Bloggie.

These videos are of me cycling this morning.

The first video, taken just after sunrise shows the camera lens was not able to focus correctly and was essentially front focusing for almost the entire video (really noticeable when the bicycle is stopped at traffic lights):
[YT]http://www.youtube.com/watch?v=CyJ6xlGvLzo[/YT]

The second video, taken around 15 minutes later, shows the camera able to focus correctly due to the increased light:
[YT]http://www.youtube.com/watch?v=kAGILSqsX3g[/YT]

My apologies if you find these videos boring - feel free to skip through most of the video and just watch a few minutes if you want.

02-10-2011, 01:48 PM	#507
jpzk Site Supporter Join Date: Jan 2009 Location: Québec Photos: Gallery \| Albums Posts: 8,251	So ... I did some AF tests using the suggested parameters by another member: ISO 3200 1/15 to 1/25s F1.8 - - lens is a FA31mm Ltd , also with the DA16-50 at F2.8 Distance to subject: 18-24 inches Tungsten lighting, dimmed to achieve the above parameters. Results: FF with both set ups (FA31 Ltd and DA16-50). I even got FF at ISO1250, with the same shutter speeds and apertures. Whether this is important or not, I had the AF assist light come on a couple of times and that did not make any difference. I tried to expose the subject so that the assist lamp would not come on, but it did, as I said, a couple of times (only). While those are not parameters I would likely use in the "real world", I was disappointed to find out that my K5 does show this low light (Tungsten) AF issue. [I can "live with that" though but am I the only one who will? Just wondering.] I contacted Pentax Imaging first, and ended up in the USA, which is fine. The fellow was trying to be helpful and mentioned that there have been a few reports of such AF issues. I asked whether they had heard anything at all regarding a fix of some sort (either firmware or hardware/return the camera ... blah, blah, blah ...) and he said "No". He also said the there has been no report from Pentax Japan, so they would not have a clue on how this is going to be handled. He suggested to contact Pentax Canada because I bought the K5 there. I called Pentax Canada: of course, they are as clueless as usual; the fellow at the repair/service department told me that I am the first person reporting such problem! No kidding!! So I suppose this is a matter of waiting for Pentax Japan for anyhting they would come up with to fix this, and perhaps I am being over confident that they will? I haven't posted any pics from the tests (yet) ... they are still on the SD card. Eventually, I will post them. For what it was worth, this was my contribution for the issue at hand. Cheers! JP Last edited by jpzk; 02-10-2011 at 02:08 PM.

Similar Threads
Thread	Thread Starter	Forum	Replies	Last Post
Help! K7 malfunctioning	samski_1	Pentax DSLR Discussion	9	07-11-2010 03:24 PM
Probably not relevant or welcome here	climbertrev	Canon, Nikon, Sony, and Other Camera Brands	4	05-24-2010 05:15 AM
camera based sensorcleaning how relevant is it.	hein	Pentax DSLR Discussion	13	09-13-2009 12:26 PM
Light fluff but news.. sort of	jeffkrol	Pentax News and Rumors	8	12-30-2008 06:30 PM

02-10-2011, 03:23 AM	#496
Tommot1965 Veteran Member Join Date: Sep 2010 Location: Perth Western Australia Photos: Gallery \| Albums Posts: 1,026	Originally posted by simico Pffff.... LOL

02-10-2011, 03:48 AM	#497
jolepp Veteran Member Join Date: Oct 2010 Location: Finland Photos: Albums Posts: 3,196	It just occurred to me that EV alone might not be a sufficient measure of problematic conditions, in addition the max. aperture of the lens used is needed as that is used for AF: the amount of light (intensity) on the AF sensor depends on these two. We might want to use an equivalent value for some arbitrary f-stop, e.g. Christine's figures of EV 2.8 and EV 2.2 with f=1.9 are equivalent to EV 3.9 and EV 3.3 for f=2.8. (The adjustment term comes from the ratio of aperture areas to which light intensity is directly proportional: 1/1.9^2 / 1/2.8^2 = (2.8/1.9)^2 = 2.1717... log2 of this gives the difference in EV units as 1.1.) The suggestion of "f=2.8 normalized" figures is implied.

02-10-2011, 05:48 AM	#498
falconeye Veteran Member Join Date: Jan 2008 Location: Munich, Alps, Germany Photos: Gallery Posts: 6,871	Originally posted by Christine Tham Forgot to mention, I had a hard time repeating the test, because the AF assist light keep coming on, even at EV3-EV4. Of course, once the AF assist light comes on, the camera is able to focus with no FF. Christine, you probably missed my earlier question: Did the AF assist light (in your camera) assist the entire focus process or just the beginning before the lens is shifted (it makes a difference if the lens wasn't already in focus). Mine uses AF assist only for the first measurement (if at all), not thru hunting. Meaning it would still miss the focus if it wasn't in focus initially. I.e., it does fine focus corrections in the dark even if starting with AF assist light on.

02-10-2011, 06:12 AM	#499
areidjr Veteran Member Join Date: Sep 2007 Location: Pittsburgh, Pa, USA Photos: Gallery Posts: 529	Originally posted by falconeye Did the AF assist light (in your camera) assist the entire focus process or just the beginning before the lens is shifted (it makes a difference if the lens wasn't already in focus). Mine uses AF assist only for the first measurement (if at all), not thru hunting. Meaning it would still miss the focus if it wasn't in focus initially. I.e., it does fine focus corrections in the dark even if starting with AF assist light on. FWIW and if you missed my earlier reply, mine behaves exactly as your. Although, on rare occasions it seem to work as expected.

02-10-2011, 09:34 AM	#500
falconeye Veteran Member Join Date: Jan 2008 Location: Munich, Alps, Germany Photos: Gallery Posts: 6,871	Originally posted by areidjr FWIW and if you missed my earlier reply, mine behaves exactly as your. Although, on rare occasions it seem to work as expected. Thanks. But Christine claims to nail focus if AF assist comes on. So, I'd like to see her response as well

02-10-2011, 10:38 AM	#501
GregK8 Site Supporter Join Date: May 2008 Location: Western New York Photos: Gallery \| Albums Posts: 614	If that all it is, why not allow the Fx button to be programmed to turn on the AF assist light?

02-10-2011, 11:03 AM	#502
Christine Tham Veteran Member Join Date: Nov 2008 Location: Sydney, Australia Photos: Gallery Posts: 2,269	Originally posted by falconeye Christine, you probably missed my earlier question: Did the AF assist light (in your camera) assist the entire focus process or just the beginning before the lens is shifted (it makes a difference if the lens wasn't already in focus). Mine uses AF assist only for the first measurement (if at all), not thru hunting. Meaning it would still miss the focus if it wasn't in focus initially. I.e., it does fine focus corrections in the dark even if starting with AF assist light on. Hi Falk Sorry I missed out on your earlier post - I tuned out of the thread and wasn't reading the posts. In response to your query: On my K-5, if the AF assist light comes on, it stays on until the camera locks focus, and the end result is a perfectly focused photo. That's why I had to ensure the AF assist light was not engaging when doing the test. PS - love the work you did on sensor blur when SR is on

02-10-2011, 11:10 AM	#503
Christine Tham Veteran Member Join Date: Nov 2008 Location: Sydney, Australia Photos: Gallery Posts: 2,269	Originally posted by jolepp It just occurred to me that EV alone might not be a sufficient measure of problematic conditions, in addition the max. aperture of the lens used is needed as that is used for AF: the amount of light (intensity) on the AF sensor depends on these two. We might want to use an equivalent value for some arbitrary f-stop, e.g. Christine's figures of EV 2.8 and EV 2.2 with f=1.9 are equivalent to EV 3.9 and EV 3.3 for f=2.8. (The adjustment term comes from the ratio of aperture areas to which light intensity is directly proportional: 1/1.9^2 / 1/2.8^2 = (2.8/1.9)^2 = 2.1717... log2 of this gives the difference in EV units as 1.1.) The suggestion of "f=2.8 normalized" figures is implied. Interesting. However, I can't seem to replicate the FF on the DA 70mm 2.4, even right down to EV 1. I also can't replicate the FF on a cheap lens - FA 28-105mm (at f3.2). I haven't tried my FA 24 at f2 but I am not sure the depth of field would be deep enough either. The error in focusing appear to be less than 5mm.

02-10-2011, 12:14 PM	#504
Gimbal Pentaxian Join Date: Apr 2007 Location: Sweden Posts: 2,106	At which distance is it best to test for FF? (Mine is so severe that I can aim at anything and it will be fuzzy unless I use live view.) I notice that many tests at near minimum range, is that the best distance to test it? I know that the DOF will be shortest but then the focusing precision will be at it's highest as well. I find 2-3meters to be really revealing om my camera. First focus using live view, then without moving the camera activate phase focus and watch the focus ring, it shouldn't move.

02-10-2011, 12:36 PM	#505
raphtze Senior Member Join Date: Dec 2010 Location: Oakland, CA Posts: 136	Originally posted by philbaum I've seen a coupla identical posts on the Nikon D7000 dslr forum on dpreview, where a camera had backfocus in normal daylight that shifted to FF in tungsten low light. Its hard for me to conceive that a mfr would ship a camera this way, but some of the similarities of this issue between Nikon and Pentax are certainly interesting. i am sure both companies wanted to get their latest and greatest pro-sumer camera out to maximize profit....but well....as we're now seeing, it probably wasn't the best. some what tangential....i wonder what the QC and testing is like when a sensor is finally developed, i would assume that pentax and nikon would be a lot more careful--well, perhaps that is what it will be like with D7K and K-5 successors. hehe one thing i know is, us forum members would probably be hired on the spot for such testing -raph

02-10-2011, 01:47 PM	#506
Smeggypants Veteran Member Join Date: Sep 2010 Posts: 1,536	Originally posted by GregK8 If that all it is, why not allow the Fx button to be programmed to turn on the AF assist light? My K20D doesn't need an assist light to focus right down to low light levels.

02-10-2011, 02:06 PM	#508
areidjr Veteran Member Join Date: Sep 2007 Location: Pittsburgh, Pa, USA Photos: Gallery Posts: 529	Originally posted by jpzk I contacted Pentax Imaging first, and ended up in the USA, which is fine. The fellow was trying to be helpful and mentioned that there have been a few reports of such AF issues. I asked whether they had heard anything at all regarding a fix of some sort (either firmware or hardware/return the camera ... blah, blah, blah ...) and he said "No". He also said the there has been no report from Pentax Japan, so they would not have a clue on how this is going to be handled. This is totally consistent with the calls I made to Pentax USA both a little over 2 weeks ago and again yesterday.

02-10-2011, 03:30 PM	#509
falconeye Veteran Member Join Date: Jan 2008 Location: Munich, Alps, Germany Photos: Gallery Posts: 6,871	Originally posted by Christine Tham In response to your query: On my K-5, if the AF assist light comes on, it stays on until the camera locks focus, and the end result is a perfectly focused photo. Now, that's interesting. Let's double check our firmware versions first, though. Unfortunately, there has been a silent firmware upgrade recently (which I did Not install). I'll post the exact release and checksum, and how you can compute your checksum asap. What OS (Mac, PC or Linux) do you use? Maybe, Pentax silently changed the way AF assist light behaves.

02-10-2011, 03:41 PM	#510
Christine Tham Veteran Member Join Date: Nov 2008 Location: Sydney, Australia Photos: Gallery Posts: 2,269	Kind of off topic - but over the years I used many different cameras, and in my experience most of them struggle to focus in low light. If you recall, I mentioned that under the conditions in which my K-5 will FF, the Canon EOS-1D Mk III refuses to lock focus. And if you like watching videos, here's an example of front focusing under low light from my Sony Bloggie. These videos are of me cycling this morning. The first video, taken just after sunrise shows the camera lens was not able to focus correctly and was essentially front focusing for almost the entire video (really noticeable when the bicycle is stopped at traffic lights): [YT]http://www.youtube.com/watch?v=CyJ6xlGvLzo[/YT] The second video, taken around 15 minutes later, shows the camera able to focus correctly due to the increased light: [YT]http://www.youtube.com/watch?v=kAGILSqsX3g[/YT] My apologies if you find these videos boring - feel free to skip through most of the video and just watch a few minutes if you want.